Applying sequential bifurcation and feed-forward simulation to reduce the prescriptive response time of a digital twin

Recent advances in digital twin research have increasingly emphasized their role in real-time decision support for production and logistics systems, yet significant methodological gaps remain in enabling timely prescriptive analytics. This study addresses these gaps by proposing and evaluating an integrated approach that accelerates a digital twin’s ability to generate effective, computationally feasible decisions under dynamic operating conditions. This research envisions the digital twin as a decision support system capable of timely, relevant, and effective decision-making. To achieve this promise, the digital twin can describe its environment, predict future events, prescribe a course of action, and then actualize this course of action in the physical system. To ensure the decision is timely, relevant, and effective, the ability to quickly prescribe courses of action becomes paramount. This research addresses the fundamental trade-off between decision speed and depth of insight. Navigating this trade-off is facilitated by effective use of models to produce timely information, rapid identification of critical factors that drive decisions, and implementation of decisions when sufficient insight is available.